• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    Method for testing an electronic air traffic control system (2), comprising the steps of: - reception by said system (2) of input data representative of the state of air traffic; - by said system: establishment of information relating to air traffic according to said input data and delivery of said information to an electronic device (6) for testing the system; - Determination by said electronic test device (6), based on the information delivered, of air traffic control instructions and supply to said system of said instructions; - receipt and processing of said instructions by said system; according to which said electronic device (6) comprises an algorithmic model (63) for automatic determination of instructions as a function of information relating to air traffic, said model having been obtained during a learning phase, implemented by computer, of a deep learning neural network, according to a set of instructions previously provided by at least one air traffic controller to the system and information relating to air traffic associated with said instructions.
    公开号:FR3082965A1
    申请号:FR1800642
    申请日:2018-06-21
    公开日:2019-12-27
    发明作者:Beatrice Pesquet Popescu；Fateh Kaakai；Frederic Barbaresco
    申请人:Thales SA；
    IPC主号:
    专利说明:

    © METHOD OF TESTING AN ELECTRONIC AIR TRAFFIC CONTROL SYSTEM, ELECTRONIC DEVICE AND ASSOCIATED PLATFORM.
    ©) Method for testing an electronic air traffic control system (2), comprising the steps of:
    - reception by said system (2) of input data representative of the state of air traffic;
    - by said system: establishment of information relating to air traffic as a function of said input data and delivery of said information to an electronic device (6) for testing the system;
    - determination by said electronic test device (6), as a function of the information delivered, of air traffic control instructions and supply to said system of said instructions;
    - reception and processing of said instructions by said system;
    according to which said electronic device (6) comprises an algorithmic model (63) for automatic determination of instructions as a function of information relating to air traffic, said model having been obtained during a learning phase, implemented by computer, a deep learning neural network, based on a set of instructions previously supplied by at least one air traffic controller to the system and associated air traffic information said instructions.
    Illllllllllllllllllllllllllllllllllllllllllll
    Method for testing an electronic air traffic control system, electronic device and associated platform
    The present invention relates to the field of electronic air traffic control systems, typically so-called electronic systems ATM, from the English "Air Traffic Management". Such a system provides the interface between on the one hand an air traffic controller, for example in charge of a given geographic area, and on the other hand the aircraft located within the geographic area or other air traffic controllers, in particular those in charge of neighboring geographic areas.
    Such a system receives data from external systems (meteorological data, aircraft flight plans, detection of radars, messages from air traffic controllers of neighboring sectors, etc.), processes this data, possibly combines them, etc., then restores, via an HMI (human-machine interface), to the air traffic controller this data or the information from the processing. The air traffic controller, on the basis of this data and information, determines air traffic control instructions (commands intended for aircraft, messages to neighboring controllers comprising given information etc.) and enters them via ΙΊΗΜ. The system then processes these commands.
    Such a system is regularly subjected to validation and integration tests. As is known, such validation and integration tests aim at verifying the proper functioning of the system, and at detecting possible bugs, first at the system manufacturer, then once the system is installed on the operating site. . They allow, for example, to verify the conformity of the behavior of the system with the specifications, both internally and in terms of its exchanges with the external interfaces, to test its performance, for example response times, and its robustness. These validation and integration tests take place with predefined or non-predefined input data. These tests must be carried out when introducing new functionalities into the electronic air traffic control system or to test the non-regression of existing functionality. They require the participation, for days, even weeks, of air traffic controllers to interact with the system, in the same way as during normal operational use, which is an obstacle to the implementation of complete and intensive tests , and in so doing the rapid deployment of technical developments in these systems.
    For example, a known feature of an air traffic control system is conflict detection, where an alert is generated if the system detects a risk of collision between two aircraft in the next n minutes (for example, n = 3 ). In such a case, by being thus notified by the alert and in view of the other data provided by the system concerning the current state of air traffic and the air environment, the controller gives aircraft audit commands and communicates if necessary with the controllers of the neighboring sectors via their respective ATM system.
    When the conflict detection method changes, for example by the use of a new conflict detection algorithm, tests are implemented, including providing the ATM system with input data which is then processed and / or displayed, their analysis by the air traffic controller, the supply of instructions by the latter, then the processing by the ATM system of these instructions, the behavior of the system being supervised to detect inconsistencies, drifts, regressions etc.
    Another functionality of an ATM system is for example the supervision of safety distances between aircraft. When the form of restitution of this functionality changes, tests must also be implemented.
    There is therefore a need to facilitate the implementation of tests of electronic air traffic control systems.
    To this end, according to a first aspect, the invention proposes a method for testing an electronic air traffic control system delivering information relating to air traffic established as a function of input data representative of the state of air traffic. received by said system, said system further receiving, and processing, in the operational phase, air traffic control instructions which are supplied to it by at least one air traffic controller, said method being characterized in that it comprises, during a test phase of said system, the steps of:
    - reception by said system of input data representative of the state of air traffic;
    - establishment, by said system, of information relating to air traffic as a function of said input data and delivery, by said system, of said information to an electronic device for testing the system;
    - determination by said electronic system test device, based on the information delivered, of air traffic control instructions and supply to said system of said instructions;
    - reception and processing of said instructions by said system;
    according to which said electronic device comprises an algorithmic model for automatic determination of instructions as a function of information relating to air traffic, said model having been obtained during a learning phase, implemented by computer, of a neural network deep learning, based on a set of instructions previously provided by at least one air traffic controller to the system and information relating to air traffic associated with said instructions.
    The invention thus makes it possible to carry out tests of electronic air traffic control systems of much higher duration and intensity, which makes it possible to accelerate and make reliable the operational putting into service of the new functionalities or more generally of the new versions. of these systems.
    In embodiments, the test method according to the invention further comprises one or more of the following characteristics:
    the test method comprises, during the test phase, the detection of nonconformity (s) of the system as a function of the behavior of the system;
    the algorithmic model of automatic determination of instructions has been learned to determine instructions specific to at least one element, as a function of a differentiation by element among several elements of the same type, during training, of the instructions previously provided by the at least one air traffic controller in the system and information relating to air traffic associated with said instructions, said element among several elements of the same type being a geographical sector among several geographical sectors and / or an air traffic controller role among several roles and / or a functionality of the electronic air traffic control system among several functionalities of said system;
    - the test process includes the steps of:
    - determination of an algorithmic constraint module adapted to identify instructions that do not comply with the rules of air controllers;
    - application of said algorithmic module to the set of instructions previously supplied by at least one air traffic controller to the system and removal of said instructions identified as non-compliant from the set used for learning the neural network;
    - application of said algorithmic module on the instructions determined by the electronic test device and not considered in the instruction test phase (s) identified as non-compliant.
    the test method comprises, in a preliminary phase to the test, the steps of:
    - collection and storage of a set of instructions previously supplied to the system by at least one air traffic controller and information relating to air traffic delivered by the system associated with said instructions;
    determination of the algorithmic model for automatic determination of instructions as a function of information relating to air traffic by learning, implemented by computer, of a neural network, as a function of said set of stored instructions and said information relating to air traffic stored and associated with said instructions.
    According to a second aspect, the present invention provides an electronic device for testing an electronic air traffic control system delivering information relating to air traffic established as a function of input data representative of the state of air traffic received by said air traffic control system. system, said system further receiving, and processing air traffic control instructions supplied to it by at least one air traffic controller, said electronic test device being characterized in that it is adapted to receive information relating to the air traffic delivered by the system and in that it comprises an algorithmic model for automatic determination of instructions as a function of the information relating to air traffic, said model having been obtained during a learning phase, implemented by computer, a deep learning neural network, based on ’A set of instructions previously provided by at least one air traffic controller to the system and associated air traffic information from said instructions.
    In embodiments, the test device according to the invention further comprises one or more of the following characteristics:
    - it is suitable for detecting non-compliance (s) of the system according to the behavior of the system;
    - the algorithmic model for automatic determination of instructions has been learned to determine instructions specific to at least one element, as a function of differentiation by element among several elements of the same type, during learning, of the instructions previously provided by at least one air traffic controller in the system and information relating to air traffic associated with said instructions, said element among several elements of the same type being a geographical sector among several geographical sectors and / or an air traffic controller role among several roles and / or a functionality of the electronic air traffic control system among several functionalities of said system.
    According to a third aspect, the present invention provides a platform for testing an electronic air traffic control system delivering air traffic information established as a function of input data representative of the state of air traffic received by said system, said system further receiving, and processing, in operational phase, air traffic control instructions which are supplied to it by at least one air traffic controller, said platform being adapted to collect and store a set of instructions previously supplied to the system by at least one air traffic controller and information relating to air traffic delivered by the system associated with said instructions;
    - determining an algorithmic model for automatic determination of instructions as a function of information relating to air traffic by learning, implemented by computer, of a neural network, as a function of said set of stored instructions and said information relating to air traffic stored and associated with said instructions;
    - obtain an electronic system test device comprising said algorithmic model for automatic instruction determination;
    - test said system by said electronic test device.
    According to a fourth aspect, the present invention provides a test platform comprising:
    an electronic air traffic control system delivering information relating to air traffic established as a function of input data representative of the state of air traffic received by said system, said system also receiving and processing traffic control instructions provided to it by at least one air traffic controller; and
    - an electronic device for testing said electronic system according to one of the claims according to the second aspect of the invention.
    These characteristics and advantages of the invention will appear on reading the description which follows, given solely by way of example, and made with reference to the accompanying drawings, in which:
    FIG. 1 represents a view of a test platform in an embodiment of the invention;
    Figure 2 is a flow chart of steps implemented in one embodiment of the invention.
    FIG. 1 represents a test platform 1 of an electronic air traffic control system. The test platform 1 comprises in the example considered an electronic air traffic control system 2, named ATM system 2 below, and an electronic test device 6.
    The ATM system 2 comprises a memory 20, an HMI block 21 and a processor 22.
    The HMI unit 21 includes for example display screens, for example touch screens, a loudspeaker system, a keyboard, a microphone, etc.
    Furthermore, the ATM system 2 is connected by telecommunications links to a set 4 of external systems 4_1, 4_2, ..., 4_n.
    These external systems include, for example, aircraft, radars, weather stations, telecommunication devices of other air traffic controllers, airport control rooms ...
    The ATM system 2 is adapted to receive, and store in its memory 20, input data (and the time stamp of this data).
    This input data includes external data, Le. which are delivered to it by the external systems 4_1, 4_2, ..., 4_n of the set 4 representing the current or future state of the airspace (or of a given sector of the airspace) comprising by example, in a non-exhaustive way:
    for each aircraft currently or shortly in the sector: 3-dimensional (3D) coordinates, type of aircraft, its heading, its speed, its flight plan; current and future weather information;
    air traffic density indices (such as number of aircraft in the area, turbulence), data defining the current and future structure of airspace (such as the presence of military no-go zones, air corridors , configuration, current and future, of the airport (open runways, wind direction, taxi lanes available ..., surveillance data (such as from primary and secondary radars, ADS-B, WAM,. ..;
    coordination messages with air traffic controllers operating on ATM systems relating to adjacent sectors.
    The ATM 2 system is further adapted to restore this data to the air traffic controller, via the HMI block 21.
    External data can also be processed (for example averaged, verified, combined with one another, analyzed, etc.) before being returned.
    The input data also includes this processed data and their time stamp.
    The ATM 2 system is further adapted to generate data internal to the ATM 2 system, representative of its current state. They are generated in particular using probes or functions installed in the ATM 2 system and include, for example, in a non-exhaustive manner:
    a message queue, shared data, logs, technical traces, a focus action on a flight tag, enlargement or zoom factor on a given area, transcription of voice conversation between a controller and a aircraft pilot, etc. The internal data is accessible to the air traffic controller operating the ATM 2 system via ΙΊΗΜ 21. In one embodiment, the input data stored in the memory 20 also includes this internal data, with the corresponding time stamping said data.
    In one embodiment, the ATM 2 system is adapted to implement advanced functions using the external ones and generating internal data. These functions are implemented for example using computer programs implementing the function, which are stored in the memory 20 and executed on the processor 22. These functions include for example the detection of conflicts, identifying a risk of collision between an aircraft and another aircraft, or the risk of an aircraft entering an area during a no-fly period in the area, etc., or the determination of a solution to resolve the conflict. The results of these functions are returned via ΙΊΗΜ 21 to the air traffic controller.
    In a known manner, an air traffic controller operating the ATM 2 system can thus become aware, via ΓΙΗΜ 21, at all times, of the current or future air situation, as a function of external data and / or as a function of internal data, including the results of advanced functions. These data and results are supplied to it via ΙΊΗΜ 21. Depending on these elements, the air traffic controller then makes decisions which it supplies to the ATM 2 system via ΙΊΗΜ 21 (in text or visual or voice form ...).
    These decisions include instructions for aircraft and / or ATM systems of adjacent sector controllers. They can thus include commands for target flight altitude, target speeds (horizontal, vertical), target heading, target climb or descent slope, etc.
    The electronic test device 6 comprises a memory 61 and a processor 62. In the memory 61, an artificial neural network 63 is notably stored.
    Figure 2 is a flow chart of steps implemented in an embodiment of the invention
    The neural network 63 is adapted for, in a preliminary phase 101_0 of obtaining the programmed network, carrying out a learning of the behavior of air traffic controller (s), from a history of input elements comprising external data from ATM systems, and comprising, in embodiments, internal data from ATM systems and / or function results, and from an output history including decisions made by air traffic controllers at seen from these respective input elements.
    This history is for example made up of the collection over several months, from the ATM 2 system, or another ATM system similar to the ATM 2 system or from several operational ATM systems, of all of these elements and their storage. In one embodiment, said storage is carried out in memory 20.
    It will be noted that this preliminary phase 101_0 of learning the neural network 63, according to the embodiments:
    - is implemented, within the electronic test device 6, using the memory 61 and the processor 62, or
    - is implemented on a specific learning platform (not shown), with its own memory and calculation resources.
    In one embodiment, for each test configuration, the preliminary phase 101_0 comprises a preparation phase 101_01 of the input and output elements identifying which are (the extracts from) those of the input and output elements useful for the test , in decision-making, as well as for example the minimum collection period.
    In a known manner, the preparation of these elements can include the segmentation of the elements collected, the detection of missing elements and aberrations, the reduction of the dimensions of the elements, the extraction of groups, the identification of causes and relations, and finally the definition of a set of drive elements comprising input drive elements and associated output elements. For each test configuration, the preliminary phase 101_0 comprises a training phase 101_02 strictly speaking of the neural network 63 from the associated input and output drive elements of the set of drive elements. A neural network per test configuration is thus determined.
    These principles of defining data sets for training, learning and the use of neural networks are well known: cf. for example :
    Tolk, A. (2015, July). The next generation of modeling & simulation: integrating big data and deep learning. In Proceedings of the Conference on Summer Computer Simulation (pp. 1-8). Society for Computer Simulation International;
    - Akerkar, R. (2014). Analytics on Big Aviation Data: Turning Data into Insights. IJCSA, 11 (3), 116-127;
    Boci, E., & Thistlethwaite, S. (2015, April). A novel big data architecture in support of ADS-B data analytic. In Integrated Communication, Navigation, and Surveillance Conference (ICNS), 2015 (pp. C1-1). IEEE;
    Bengio, Y. (2009). Learning deep architectures for Al. Foundations and trends® in Machine Learning, 2 (1), 1-127.
    The result of training phase 101_02 is the delivery 101_1 of a trained neural network 63, also called an air traffic controller (algorithmic) model. This model is then saved in the electronic device 6, which is then able to be used in the test phase 101_2 to test the ATM system 2.
    In embodiments, the input and output elements are structured by aerial geographic sector, the training of the neural network is then also differentiated by sector, and the behavior of the algorithmic air traffic controller model 63 obtained in step 101_1 is specific to each sector.
    Similarly, in embodiments, the sets of drive elements are structured by technical functionality of the ATM system (for example conflict detection or flight altitude commands or coordination between controllers) or by the specific role of a controller. air (for example, command role or role of planning and exchange with adjacent sectors). In such a case, the resulting trained model is then specific to a role or a functionality. In addition, mixed modes can be generated, specific to at least two aspects among the aspects of sector, functionality and role.
    Any type of artificial neural network can be used. For example, a deep learning network, English "deep learning network", a CNN convolutional neural network, English "Convolutional Neural Network" are used. The number of input nodes will be chosen equal to the number of input elements and the number of output nodes will be chosen equal to the number of output elements for each test configuration.
    During a test phase 101_2 of the ATM 2 system, the following steps are implemented and repeated:
    reception by the ATM 2 system of external data representative of the current state of air traffic;
    - establishment, by said ATM 2 system, of information relating to air traffic as a function of said data, comprising at least some of the external data, as well as of updated internal data, for example results of functions of the ATM 2 system and delivery , by said system 2, via ΙΊΗΜ 21, information to the electronic test device 6;
    - determination by the neural network 63 of the electronic test device 6, as a function of the information delivered, of output elements comprising air traffic control decisions; these decisions include instructions for aircraft and / or ATM systems of adjacent sector controllers;
    - And supply by the test device 6 via ΙΊΗΜ 21 to said ATM system 2 of said output elements;
    reception and processing of said output elements by said system 2, comprising the transmission of decisions to aircraft and / or to ATM systems of controllers of adjacent sectors.
    The nonconformities of the system are detected during the testing phase, by analyzing the behavior of the system 2. This detection is done for example by the test device 6 or any other means.
    Several types of test aimed at distinct purposes can be implemented in the test phase 101_2 exist.
    One type of test is for example an extensive verification test campaign, based on actual input data (ie internal data, or even external data) of the ATM2 system, from which test sets of input data of the controller model are generated randomly (some within the permitted ranges, others outside the permitted ranges to test the robustness). The decisions of controller model 63 will be consistent with the behavior of the human controller as learned.
    Another type of test phase is for example a non-regression test campaign: in such a case, using the test device 6 based on the controller model 63 generated before the evolution of the ATM system 2, the system ATM2 is tested. taking into account the evolution by providing the ATM 2 system with external input data equal to that which enabled this controller model to be learned previously. If there is no injection into the ATM 2 system of functional deviation from the specification (software bug, side effect of a technical change such as hardware update ...), it is expected to have exactly the same exit decisions issued by the test device 6. Otherwise, there is potentially a regression in the ATM 2 system and additional investigations must be carried out.
    Another type of test phase is for example an endurance test campaign: such a campaign is similar to a non-regression campaign except that the objective is different. The purpose of an endurance test campaign is to exploit the previous status of non-regression tests to build a service experience, which will be:
    - performed at an early stage in the ATM 2 system life cycle, ie before the operational transition, carried out in an accelerated manner by duplicating numerous instances of the ATM 2 system to be validated and numerous instances of controller models and data sets d 'test entry.
    An ATM 2 system evolving regularly, updates both functional (introduction of new functions or modified functions) and technical (changes in equipment, operating system etc.) are to be taken into account in the tests. In such a situation, in one embodiment, the algorithmic model 63 corresponding to the ATM system 2 before update is completed to take these updates into account. Thus with reference to FIG. 2, in a step 102_0, input elements comprising internal data from ATM systems, external data from ATM systems, and output elements comprising the decisions taken by the air traffic controllers in view of these respective input elements, are recorded and stored during validation sessions carried out by air traffic controllers on a test platform representative of the operational platform relating to the part which is updated from the ATM 2 system. step 101_0 of obtaining a programmed neural network is then implemented on the basis of these input and output elements, and gives rise to the delivery of an air network controller model targeted on the part of the ATM 2 system which is updated. In step 101_1, a combination of the algorithmic model corresponding to the ATM 2 system before updating and the algorithmic model of the ATM 2 system targeted on the updated aspects is produced (for example, in embodiments, by concatenation ), thus making it possible to deliver a complete algorithmic model corresponding to the updated ATM 2 system. Furthermore, in one embodiment, in a security step 100_0, of the rules, principles, constraints, conditions and prohibitions implemented by the air traffic controllers in the application of their profession (corresponding for example to the ICAO standards as defined in document 4444) are formalized in algorithmic form.
    For example, it is among these rules that:
    a1 / an air traffic controller cannot provide orders for aircraft outside the sector for which he is responsible, a2 / except for certain exceptions defined precisely by the conditions Condl, Cond2, Cond3;
    - a3 / in a given situation corresponding to a given speed and altitude of an aircraft, a controlled change in altitude of flight level must be less than a given threshold depending on said speed and altitudes.
    The resulting security algorithm is, in one embodiment, implemented in a step 100_1 on the input and output drive elements associated with the set of drive elements prior to the construction of a model. of air traffic controller, which makes it possible to detect elements which do not conform to standard practice, and then either to remove them from the set of training elements, or to assign them to a “bad practice” class allowing the model to it would be better to learn the behavior of a controller in accordance with "good practice".
    In one embodiment, in a step 100_1, these rules, principles, constraints, conditions and prohibitions implemented by the air traffic controllers in the application of their profession are also provided as training data to a network of artificial neurons, and, at the end of the learning phase, a model including these rules is delivered, hereinafter called the security model. In a step 101_3, each decision then taken by the controller model 63 during the test phase 101_2 is supplied to this security model, which either validates the decision as conforming to good practice (notably valid that it is in the dynamic range acceptable for exit decisions), or invalidate the decision which is then not taken into account in the context of the tests and comes, for example, to enrich the class of "bad practice", which has the effect of strengthening the air traffic controller model.
    The present invention thus makes it possible to obtain an electronic test device 6 based on an air controller model 63 learned by neural network.
    In the test phases, the air traffic controller model 63 obtained can be duplicated and each copy can be installed in the same electronic device 6 or installed in respective test devices similar to the device 6; this makes it possible to increase the number and extent of tests carried out in parallel on the ATM 2 system, or even on instances which are also duplicated (on the computer cloud, for example) of the ATM 2 system. embodiments of the test phase implemented using these devices, the input data can be distributed between the different models or devices, each model being assigned to a specific specific sector, or to a time period separate (peak hours, excluding peak hours, weekly, monthly periods, etc.)
    In another embodiment, the neural network model 63 is produced in the form of a programmable logic component, such as a GPU card (de ’ang aïsGraphics Processing Unit) or multi-GPU.
    权利要求:
    Claims (10)
    [1" id="c-fr-0001]
    1. - Method for testing an electronic air traffic control system (2) delivering information relating to air traffic established as a function of input data representative of the state of air traffic received by said system, said system in addition to receiving, and processing, in the operational phase, air traffic control instructions which are supplied to it by at least one air traffic controller, said method being characterized in that it comprises, during a test phase of said system, the stages of:
    reception by said system (2) of input data representative of the state of air traffic;
    - establishment, by said system, of information relating to air traffic as a function of said input data and delivery, by said system, of said information to an electronic device (6) for testing the system;
    - determination by said electronic system test device (6), according to the information delivered, of air traffic control instructions and supply to said system of said instructions;
    - reception and processing of said instructions by said system;
    according to which said electronic device (6) comprises an algorithmic model (63) for automatic determination of instructions as a function of information relating to air traffic, said model having been obtained during a learning phase, implemented by computer, a deep learning neural network, based on a set of instructions previously supplied by at least one air traffic controller to the system and associated air traffic information said instructions.
    [2" id="c-fr-0002]
    2. - Method for testing an electronic air traffic control system (2) according to claim 1, comprising during the test phase, detecting non-compliance (s) of the system as a function of the behavior of the system.
    [3" id="c-fr-0003]
    3. - Method for testing an electronic air traffic control system (2) according to claim 1 or 2, according to which the algorithmic model (63) for automatic instruction determination has been learned to determine instructions specific to the at least one element, as a function of differentiation by element among several elements of the same type, during learning, of the instructions previously supplied by at least one air traffic controller to the system and of the information relating to air traffic associated with said instructions, said element among several elements of the same type being a geographical sector among several geographical sectors and / or an air traffic controller role among several roles and / or a functionality of the air traffic control electronic system (2) among several functionalities of said system.
    [4" id="c-fr-0004]
    4. - Method for testing an electronic air traffic control system according to one of claims 1 to 3, comprising the steps of:
    - determination of an algorithmic constraint module adapted to identify instructions that do not comply with the rules of air controllers;
    - application of said algorithmic module to the set of instructions previously supplied by at least one air traffic controller to the system and removal of said instructions identified as non-compliant from the set used for learning the neural network;
    - application of said algorithmic module on the instructions determined by the electronic test device and not considered in the instruction test phase (s) identified as non-compliant.
    [5" id="c-fr-0005]
    5. - Method for testing an electronic air traffic control system (2) according to one of claims 1 to 4, comprising, in a phase preliminary to the test, the steps of:
    - collection and storage of a set of instructions previously supplied to the system by at least one air traffic controller and information relating to air traffic delivered by the system associated with said instructions;
    determination of the algorithmic model (63) for automatic determination of instructions as a function of information relating to air traffic by learning, implemented by computer, of a neural network (63), as a function of said set of instructions stored and said stored air traffic information associated with said instructions.
    [6" id="c-fr-0006]
    6. - Electronic device (6) for testing an electronic air traffic control system (2) delivering information relating to air traffic established as a function of input data representative of the state of air traffic received by said system , said system further receiving, and processing air traffic control instructions supplied to it by at least one air traffic controller, said electronic test device (6) being characterized in that it is adapted to receive information relating to air traffic delivered by the system (2) and in that it comprises an algorithmic model (63) for automatic determination of instructions as a function of information relating to air traffic, said model having been obtained during a phase learning, computer implementation of a deep learning neural network, based on a set of pre instructions previously provided by at least one air traffic controller to the system and associated air traffic information.
    [7" id="c-fr-0007]
    7, - Electronic test device (6) according to claim 6, adapted to detect the nonconformity (s) of the system according to the behavior of the system.
    [8" id="c-fr-0008]
    8, - An electronic test device (6) according to claim 6 or 7, in which the algorithmic model (63) for automatic determination of instructions has been learned to determine instructions specific to at least one element, according to a differentiation by element among several elements of the same type, during learning, of the instructions previously supplied by at least one air traffic controller to the system and of information relating to air traffic associated with said instructions, said element among several elements of the same type being a sector geographic among several geographic sectors and / or an air traffic controller role among several roles and / or a functionality of the electronic air traffic control system (2) among several functionalities of said system.
    [9" id="c-fr-0009]
    9, - Platform for testing an electronic air traffic control system (2) delivering information relating to air traffic established as a function of input data representative of the state of air traffic received by said system, said system system further receiving, and processing, in the operational phase, air traffic control instructions supplied to it by at least one air traffic controller, said platform being adapted to collect and store a set of instructions previously supplied to the system by at least one air traffic controller and information relating to air traffic delivered by the system associated with said instructions;
    determining an algorithmic model (63) for automatic determination of instructions as a function of information relating to air traffic by learning, implemented by computer, of a neural network (63), as a function of said set of instructions stored and said air traffic information stored and associated with said instructions;
    - obtain an electronic device (6) for testing the system (2) comprising said algorithmic model (63) for automatic determination of instructions;
    - testing said system (2) by said electronic test device (6).
    [10" id="c-fr-0010]
    10.- Test platform (1) including:
    an electronic air traffic control system (2) delivering
    5 air traffic information established on the basis of input data representative of the state of air traffic received by said system, said system also receiving and processing air traffic control instructions which are supplied to it by at least one air traffic controller; and an electronic device for testing said electronic system according to one of claims 10 to 8.
    类似技术:
    公开号 | 公开日 | 专利标题
    EP3588387A1|2020-01-01|Method for testing an electronic air traffic control system, associated electronic device and platform
    US20170039890A1|2017-02-09|Augmenting transport services using driver profiling
    FR3052273A1|2017-12-08|PREDICTION OF TROUBLES IN AN AIRCRAFT
    FR2963448A1|2012-02-03|METHOD AND SYSTEM FOR ANALYSIS OF FLIGHT DATA RECORDED DURING THE FLIGHT OF AN AIRCRAFT.
    US20160370332A1|2016-12-22|Generating fine resolution air pollution estimates
    US20190213502A1|2019-07-11|Creating a signature for identifying a particular machine learning model
    FR3037679A1|2016-12-23|
    WO2021232229A1|2021-11-25|Virtual scene generation method and apparatus, computer device and storage medium
    WO2011032781A1|2011-03-24|Method and apparatus for configuring a motion simulation device
    EP3588386A1|2020-01-01|Training and/or assistance platform for air management via air traffic management electronic system, associated method
    WO2021122397A1|2021-06-24|Method and device for supervising a traffic control system
    CN103942403A|2014-07-23|Method and device for screening mass variables
    EP2939902A1|2015-11-04|Method and device for determining a railway itinerary
    Pötzi et al.2018|An event-based verification scheme for the real-time flare detection system at Kanzelhöhe observatory
    FR3038751A1|2017-01-13|METHOD FOR INTEGRATING A CONSTRAINED ROAD OPTIMIZATION APPLICATION IN AN OPEN ARCHITECTURE AIRCRAFT SYSTEM OF CLIENT-TYPE SERVER
    FR3082330A1|2019-12-13|AERONAUTICAL CYBERSECURITY
    FR2973131A1|2012-09-28|METHOD AND DEVICE FOR DETECTING INCOMPATIBILITIES OF LOGIC INTERFACES OF ONBOARD SYSTEM EQUIPMENT
    RU122796U1|2012-12-10|DECISION MAKING SUPPORT SYSTEM FOR PSYCHOLOGICAL AND PEDAGOGICAL TESTING
    US20210118117A1|2021-04-22|Predicting and correcting vegetation state
    Haq et al.2021|Can Offline Testing of Deep Neural Networks Replace Their Online Testing?
    Wang et al.2017|A large scale neural network model for crash prediction in urban road networks
    JP6681965B2|2020-04-15|Apparatus and method for extracting learning target image for autonomous driving
    US10769768B1|2020-09-08|Apparatus and method for providing application service using satellite image
    Clark et al.2009|Advanced technologies for acoustic monitoring of bird populations
    WO2019229174A1|2019-12-05|Method for controlling the malfunctions of a piece of equipment and associated devices
    同族专利:
    公开号 | 公开日
    SG10201905716VA|2020-01-30|
    FR3082965B1|2022-01-28|
    AU2019204291A1|2020-01-16|
    CN110634329A|2019-12-31|
    JP2020024678A|2020-02-13|
    TW202001559A|2020-01-01|
    KR20190143832A|2019-12-31|
    RU2019119152A|2020-12-21|
    US20190391909A1|2019-12-26|
    BR102019012735A2|2019-12-24|
    EP3588387A1|2020-01-01|
    CA3047430A1|2019-12-21|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    US20160314692A1|2015-04-22|2016-10-27|The Boeing Company|Data driven airplane intent inferencing|
    WO2020097230A1|2018-11-06|2020-05-14|Vianair Inc.|Airspace information modeling and design|
    CN111047915B|2019-12-13|2020-11-27|中国科学院深圳先进技术研究院|Parking space allocation method and device and terminal equipment|
    法律状态:
    2019-07-01| PLFP| Fee payment|Year of fee payment: 2 |
    2019-12-27| PLSC| Publication of the preliminary search report|Effective date: 20191227 |
    2020-06-30| PLFP| Fee payment|Year of fee payment: 3 |
    2021-06-30| PLFP| Fee payment|Year of fee payment: 4 |
    优先权:
    申请号 | 申请日 | 专利标题
    FR1800642A|FR3082965B1|2018-06-21|2018-06-21|METHOD FOR TESTING AN ELECTRONIC AIR TRAFFIC CONTROL SYSTEM, ELECTRONIC DEVICE AND ASSOCIATED PLATFORM|
    FR1800642|2018-06-21|FR1800642A| FR3082965B1|2018-06-21|2018-06-21|METHOD FOR TESTING AN ELECTRONIC AIR TRAFFIC CONTROL SYSTEM, ELECTRONIC DEVICE AND ASSOCIATED PLATFORM|
    AU2019204291A| AU2019204291A1|2018-06-21|2019-06-19|Method for testing air traffic management electronic system, associated electronic device and platform|
    BR102019012735A| BR102019012735A2|2018-06-21|2019-06-19|test method for an electronic air traffic control system, electronic test device and test platform|
    CA3047430A| CA3047430A1|2018-06-21|2019-06-19|Test process for an electronic air traffic control system, electronic device and associated platform|
    JP2019114553A| JP2020024678A|2018-06-21|2019-06-20|Method for testing air traffic control electronic system, related device, and platform|
    US16/447,612| US20190391909A1|2018-06-21|2019-06-20|Method for testing air traffic management electronic system, associated electronic device and platform|
    SG10201905716VA| SG10201905716VA|2018-06-21|2019-06-20|Method for testing air traffic management electronic system, associated electronic device and platform|
    EP19181533.1A| EP3588387A1|2018-06-21|2019-06-20|Method for testing an electronic air traffic control system, associated electronic device and platform|
    TW108121470A| TW202001559A|2018-06-21|2019-06-20|Method for testing air traffic management electronic system, associated electronic device and platform|
    RU2019119152A| RU2019119152A|2018-06-21|2019-06-20|METHOD FOR TESTING ELECTRONIC SYSTEM OF AIR TRAFFIC ORGANIZATION, ASSOCIATED ELECTRONIC DEVICE AND PLATFORM|
    KR1020190074160A| KR20190143832A|2018-06-21|2019-06-21|Method for testing air traffic management electronic system, associated electronic device and platform|
    CN201910541681.5A| CN110634329A|2018-06-21|2019-06-21|Test method of air traffic management electronic system, related electronic equipment and platform|
    [返回顶部]